My USB drive suddenly stopped opening, and Windows now says it needs to be formatted. It contains important files that I can’t afford to lose. How can I recover the data or repair the USB drive without formatting it?
Don’t format the USB yet. That Windows message usually means the file system is corrupted or unreadable, not necessarily that your photos, documents, and videos are gone. It can happen after a power loss, an unsafe disconnect, or unplugging the drive while it was still writing data.
I’ve dealt with this more than once, and the safest approach has been to recover the files first, then format the drive afterward. Windows may offer formatting as the fix, but doing that now can make recovery more difficult.
I use Disk Drill for this. Before scanning, I’d make a byte-to-byte backup of the USB if possible. A format warning can sometimes show up when the drive has bad blocks or is starting to fail, so repeatedly scanning the physical device may put more stress on it. Disk Drill can create a complete disk image, which you can then load through “Attach Disk Image” and scan instead of working directly from the USB.
Select the USB or attached disk image, then click “Search for lost data.” Let the Universal Scan finish if you can. It can look for files by their signatures rather than relying entirely on the damaged file system, so it may still find data even when Windows sees the drive as RAW. The software recognizes roughly 400 file types, including photos, documents, and video files.
You can open “Review found items” while the scan is still running. The category filters help narrow things down, especially if you only need pictures or documents. Use the eye icon to preview important files before recovering them. If a file displays correctly in the preview, that’s a good sign it can be recovered intact.
Choose what you want, click “Recover,” and save everything to a different drive, such as your computer’s internal storage. Do not recover files back onto the damaged USB because that could overwrite other data you’re still trying to retrieve.
Once the recovered files are safely stored and you’ve checked that they open properly, you can format the USB. exFAT is a practical choice if you need broad compatibility, while NTFS may make more sense if the drive will mainly be used with Windows.
The order matters here: recover first, format second. It takes a little longer, but it gives you a much better chance of getting the files back before repairing the drive.
If the USB keeps disconnecting, shows the wrong capacity, or makes the computer freeze, stop testing it with software. That points more toward hardware failure, and repeated scans or CHKDSK can make recovery worse. @zerostream3493’s image-first approach makes sense only if the drive stays connected and reports its full size. Otherwise, the safest option for truly irreplaceable files is a professional recovery service before attempting Disk Drill or any repair command.
Do not click Format, initialize the disk, create a new volume, or run CHKDSK on the original USB. CHKDSK is a repair tool, not a recovery tool, and its changes can remove damaged directory entries that recovery software might otherwise use.
Open Disk Management and inspect the USB without changing anything. Pay attention to the reported capacity and status:
- If it shows the correct size but the partition is RAW, the problem is probably the file system.
- If it shows a normal partition with no drive letter, assigning a letter may be enough.
- If it says “No Media,” reports 0 bytes, shows a wildly incorrect capacity, or repeatedly vanishes, software recovery is unlikely to solve it.
There is another possibility people miss: Windows may be asking to format a perfectly readable drive because it uses a file system Windows does not support. If the USB was recently used with a Mac, Linux machine, encrypted computer, camera, recorder, or another device, try reading it on that original device or operating system first. Do not let that device “prepare,” “repair,” or reformat it either.
If Disk Management shows the correct capacity and the connection remains stable, I agree with @zerostream3493 about making an image before recovery. Scan the image with Disk Drill or another recovery program rather than repeatedly scanning the USB itself. Be prepared for signature-based recovery to return files without their original names or folder structure, especially if the file system metadata is badly damaged. Previewing a file is useful, but open the recovered copy afterward to make sure it is actually complete.
Save recovered data to a separate internal or external drive. Once you have two copies of the important files and have checked them, you can format and test the USB. If it produces errors again, retire it rather than trusting it with the only copy of anything important. If the files are irreplaceable and the drive has any of the hardware symptoms @grace mentioned, stop before experimenting and get a professional assessment.
Whether the USB was protected with BitLocker or another encryption tool matters before you scan it. A raw recovery scan cannot identify normal photos or documents inside encrypted data unless the volume can be unlocked. If encryption was enabled, find the password or recovery key now. A scan that finds nothing in that situation does not automatically mean the files are gone.
For an ordinary unencrypted flash drive, I’d do only a couple of harmless connection checks: plug it directly into a different USB port, avoid hubs, and try one other computer. Cancel any format or repair prompt. If it suddenly opens, copy the irreplaceable files first. Don’t spend time sorting folders or trying to copy the entire drive in one shot. Start with the files you cannot replace, because a failing drive may not stay readable for long.
If it remains RAW but shows the correct capacity and stays connected, make an image and work from that. Recovery software can scan the image without repeatedly reading the USB. Keep the original untouched in case the first recovery attempt gives poor results. You may get files back with generic names and no folders, so search by file type and date rather than expecting the old directory layout to survive.
I agree with @ritatheraven about leaving CHKDSK alone at this stage. Assigning a missing drive letter is fairly low risk when Disk Management shows a healthy recognized partition, but CHKDSK is a different matter because it writes changes to the file system. That is something to consider after recovery, not as a way to recover the only copy.
If the drive reports zero bytes, “No Media,” an obviously wrong size, or disappears whenever it is accessed, stop the software attempts. The controller or flash memory may be failing, and no recovery program can fix a device that no longer provides readable sectors. For genuinely irreplaceable data, that is the point where paying for a professional assessment makes more sense than trying random commands from videos.
Whatever you recover should go to another physical drive. Open a sample of the recovered documents, play videos past the first few seconds, and check that larger files are complete. Only after verifying the copies should you format the USB, and I would retire it if it caused this problem without an obvious explanation. A reformatted drive can appear fine for a while and still fail again.
Don’t start creating an image until you check the destination drive. It needs free space roughly equal to the USB’s full capacity, and FAT32 cannot hold a single image file larger than 4 GB. Use an NTFS or exFAT destination so the copy does not fail halfway through.
Once the image exists, disconnect the original USB and scan only the image. That keeps recovery attempts from repeatedly reading a drive that may be deteriorating, while preserving the original for another method or professional recovery.
Realistically, if this was a cheap no-name USB stick, don’t expect a clean recovery with folders and filenames intact. Those controllers fail ugly, and signature scans usually hand you back a pile of unnamed files. The encryption warning from @quantumkernel_34 is the point most people skip past though, so check that first before you waste hours imaging a drive that scans as noise. Where I’d gently disagree with the imaging advice is for a small drive that still reads fine at the right capacity. If it mounts long enough to grab the files, just copy the important stuff off immediately and skip the whole image step. Disk Drill is fine for the RAW case where the drive stays connected, but it can’t do anything the hardware won’t let it read, so if the stick keeps dropping off, no software order of operations saves you.
Assume the USB may only stay readable once, so don’t waste that window opening files or copying folders in order. Cancel the format prompt, copy the smallest irreplaceable files to another drive first, and stop if the drive disconnects or reports the wrong size.
Expect that even a ‘successful’ recovery from a corrupted stick may hand you back files that are mostly complete but a few that quietly aren’t. That is normal, not a sign you did something wrong. Photos come back with green bars across half the image, videos stop playing partway, Office docs open but throw a repair prompt. So budget time to actually verify each irreplaceable file, not just trust that the count looks right.
The encryption check @quantumkernel_34 raised is the one I’d move to the very top. It gets skipped because people assume ‘I never set up BitLocker,’ but Windows on some laptops turns on device encryption by default now, and USB sticks formatted through certain work laptops can inherit that. If the drive was ever plugged into a managed or encrypted machine, a raw scan will just see noise and you’ll waste an evening imaging garbage. Confirm that before anything else.
On the imaging debate between @binary_optic_wizard and the earlier folks, I lean toward image-first, but with a condition nobody spelled out clearly. Imaging is worth it when the drive stays connected but reads slow or throws read errors, because you capture whatever it gives you once and stop touching the failing hardware. If the stick mounts clean at the right size and just needs a drive letter or shows RAW without any dropouts, skip the image and copy your files off right then. Making a full byte-for-byte image of a healthy 64GB stick before you’ve even tried grabbing the important 200MB is over-engineering it.
Disk Drill is a reasonable pick for the RAW-but-stable case, and its ability to scan an attached image instead of the live drive is genuinely the useful part here. Just don’t expect it to do magic on a controller that keeps dropping the connection. No software reorders its steps around hardware that won’t stay online.
One practical thing that trips people up with the imaging route @neonsprite mentioned: your destination needs space for the whole drive capacity, not the used portion. A near-empty 128GB stick still images to a 128GB file because the tool copies every sector including the empty ones. People start the image, walk away, and come back to a failed job because the target ran out of room. Check that first.
If the money side matters to you, decide early. Professional recovery on a dead controller can run into the hundreds, and for a cheap promotional stick full of stuff you could re-download, that math rarely works. For the only copy of something like family photos, it does. Sort your files by ‘can I get this back another way’ before you sort them by anything else, because that answer decides whether you keep experimenting or stop and pay someone.